Abstract
Background: We have previously shown that conserving six N-linked glycosylation sites of the human factor VIII (FVIII) B-domain sequences in a synthetic linker, V3 (31aa/N6), increased FVIII expression by 3-fold over B-domain deleted FVIII (BDD-FVIII) following AAV mediated gene transfer in mice (McIntosh et. al. Blood 2013). These N-linked glycosylation consensus sequences (N-X-T/S) are highly conserved across different species but their role in the context of transgenic FVIII expression following gene transfer remains unknown. We, therefore, undertook systemic B-domain linker engineering studies to elucidate the mechanism by which the V3-linker improves FVIII expression following AAV mediated gene transfer.
Methods: A series of FVIII variants were made by modifying the B-domain linker size and amino acid sequence and then cloned into AAV or lentiviral vectors. The latter were used for making stable FVIII producer cell lines to facilitate biochemical characterization including glycome profile using MALDI-TOF mass spectrometry. The AAV vectors, in which expression of the FVIII gene was directed by a small synthetic liver specific promoter (HLP) were used to assess transgene expression in-vitro and in murine models.
Results: Purified recombinant FVIII-V3 had the same biological activity as that of full-length and BDD FVIII protein concentrates when assessed by thrombin generation and vWF/FVIII binding affinity assays (Dissociation constant (KD) <1nM) suggesting that V3 linker did not interfere with the potency of FVIII. Replacement of the asparagine residues with alanine (FVIII-V5) within the six N-linked glycosylation motifs resulted in a >20 reduction in glycan structures but the expression profile and biological properties of FVIII-V5 were comparable to those of FVIII-V3. In contrast, FVIII-V1 (44aa), in which the N-linked glycosylation triplets in FVIII-V3 were flanked by two additional amino acids that are normally adjacent to these moieties in the native FVIII B-domain, was expressed at 3-fold low levels when compared to FVIII-V3 and V5 in mice. Similarly, randomly scrambling the amino acid sequence of the V3-linker (FVIII-V2) without modifying size resulted in a 7-fold decrease in FVIII expression levels in mice. Notably, FVIII-V3 and V5 mRNA levels were 4-fold higher than the levels observed with BDD-FVIII and FVIII-V1 variants. Deletion analysis of the V3 B-domain linker identified a shorter form (FVIII-N2 19aa/N2) that exhibited similar levels of FVIII secretion to those observed with FVIII-V3.
Conclusion: We have critically evaluated the role of the FVIII B domain V3 linker on human FVIII expression in human cell lines and mice following AAV mediated gene transfer. In all cases, the FVIII-V3 variant performed better than the BDD-FVIII, in large part due to an increase in FVIII-V3 mRNA levels. Although the FVIII-V3 molecule is fully glycosylated, our studies show that glycan content did not significantly influence FVIII expression but increasing size and modifying amino-acid content of the linker, markedly reduced FVIII expression following AAV mediated gene transfer in mice. Therefore, the results presented here support the use of FVIII-V3 for gene therapy for hemophilia A in our on-going investigator lead Phase 1 trial.
McIntosh:Freeline: Consultancy, Equity Ownership; BioMarin: Patents & Royalties. Dane:Freeeline: Employment, Equity Ownership. Tuddenham:BioMarin: Consultancy, Patents & Royalties; Freeline: Consultancy. Nathwani:Freeline: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties; BioMarin: Consultancy, Patents & Royalties; UniQure: Patents & Royalties.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal